\documentclass[a4paper, 12pt]{article}
\usepackage{setspace}
\usepackage{fullpage}
\usepackage{amsmath}
\usepackage[protrusion=true,expansion=true]{microtype}
\usepackage{graphicx}
\usepackage{wrapfig}
\usepackage{csquotes}
\usepackage{booktabs}
\usepackage{morefloats}
\usepackage{caption}
\usepackage{threeparttable}
\usepackage{pgf, tikz}
\usetikzlibrary{arrows, automata}
\usepackage{etex}
\usepackage[round]{natbib}
\usepackage{tgtermes}
\usepackage[T1]{fontenc}
\linespread{2}
\usepackage{url}
\usepackage{color}
\usetikzlibrary{positioning,shapes.geometric,arrows,decorations.markings} 
\usepackage{subcaption}
\usepackage{hyperref}
\hypersetup{colorlinks=true,linkcolor=black,citecolor=[rgb]{0.2235294, 0.3372549, 0.5098039}}
\newtheorem{hypothesis}{Hypothesis}
\usepackage{lscape}
\usepackage{xcolor}
\usepackage{import}
\usepackage{verbatim}
\usepackage{amsmath}
\usepackage{titling}
\usepackage{titletoc}
\usepackage{enumitem}
\usepackage{amssymb}
\usepackage{blindtext}
\usepackage{titlesec}
\setcounter{tocdepth}{2}
\usepackage{pdflscape}
\usepackage{afterpage}
\usepackage[symbol]{footmisc}
\newcommand{\sym}[1]{\rlap{#1}}
\makeatletter
\let\estinput=\@@input

\newcommand{\estauto}[3]{
	\vspace{.75ex}{
		\begin{tabular}{@{\hskip\tabcolsep\extracolsep\fill}l*{#2}{#3}} 
			\toprule
			\estinput{#1}
			\bottomrule
			\addlinespace[.75ex]
		\end{tabular}
	}
}		

\newcommand{\specialcell}[2][c]{
	\begin{tabular}[#1]{@{}c@{}}#2\end{tabular}
}

\newcommand{\Figtext}[1]{%
 \begin{tablenotes}[para,flushleft]
 \hspace{0pt}
 \hangindent=0.25em
 #1
 \end{tablenotes}
}

\newcommand{\Spec}{Specifications estimated using OLS including randomization block fixed effects, enumerator fixed effects, and predetermined covariates defined at the electoral district and individual respondent levels. Weights: \textit{Obs}: observations in electoral district; \textit{Reg}: registered voters in electoral district. Standard errors clustered by electoral district in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.}

\newcommand{\SpecCand}{Specifications estimated using OLS including randomization block fixed effects, enumerator fixed effects, and predetermined covariates defined at the electoral district, individual respondent, and candidate levels. Weights: \textit{Obs}: observations in electoral district; \textit{Reg}: registered voters in electoral district. Standard errors clustered by electoral district in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.}

\newcommand{\SpecWgt}{Specifications estimated using OLS including randomization block fixed effects, enumerator fixed effects, and predetermined covariates defined at the electoral district and individual respondent levels. Weights: \textit{Dem} weights observations to be representative at the electoral district-level with respect to gender and education. Standard errors clustered by electoral district in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.}

\newcommand{\SpecMixed}{All specifications estimated using OLS including randomization block fixed effects. Panel 1 adds enumerator fixed effects, and predetermined covariates defined at the electoral district, individual respondent, and candidate levels. Weights: \textit{Obs}: observations in electoral district; \textit{Reg}: registered voters in electoral district. Panel 2 adds predetermined covariates defined at the electoral district, polling station, and candidate levels. Weights: \textit{PS}: polling stations in electoral district; \textit{Reg}: registered voters at polling station. Standard errors clustered by electoral district in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.}

\newcommand{\SpecMixedPanel}{All specifications estimated using OLS including randomization block fixed effects. Panel 1 adds enumerator fixed effects, and predetermined covariates defined at the electoral district, individual respondent, and candidate levels. Weights: \textit{Obs}: observations in electoral district; \textit{Reg}: registered voters in electoral district. Standard errors clustered by electoral district in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.}

\newcommand{\SpecBalance}{All specifications are estimated using OLS and include randomization block fixed effects. Standard errors clustered by electoral district in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01.}

\newcommand{\SpecCompliers}{\Figtext{\footnotesize \textit{Notes}: Each variable is standardized. Panel A uses citizen survey to construct policy alignment; Panel B uses candidate survey (see Appendix A.6.3). Mean of each variable computed for all candidates (All); compliers (C); always-takers (AT); never-takers (NT) following \citet{abadie2003semiparametric} and \citet{angrist2013extrapolate}. Columns 5 and 6 provide bootstrapped $p$-values comparing C to AT and NT.}}


\begin{document}

\begin{center} \large \singlespacing
\textbf{Who Debates, Who Wins? At-Scale Experimental Evidence \\ on the Supply of Policy Information in a Liberian Election}

\

\normalsize  Jeremy Bowles \& Horacio Larreguy

\

\textbf{Replication of Tables and Figures}
\end{center}

\addtocontents{toc}{\protect\setcounter{tocdepth}{3}}

\tableofcontents

\clearpage

\section{Main paper}

\begin{table}[!!htpb] \centering
\begin{threeparttable}\footnotesize
\caption{Candidate characteristics \label{table_1}} \medskip
\estauto{"Tables/Table 1.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Mean of responses among incumbents, challengers, and other candidates to our candidate survey. \textit{Age}: age in years; \textit{University educated}: completed university; \textit{Ran before}: ran for office at least once before; \textit{Govt. job before}: previously held non-elected government job; \textit{NGO job before}: worked for an NGO before; \textit{Advocacy experience}: worked on an advocacy campaign before; \textit{Campaign expenditure}: self-reported campaign spending in USD; \textit{Radio station}: either owns or manages a radio station. }
\end{threeparttable}
\end{table}

\begin{table}[!!htpb]\centering
\begin{threeparttable}\footnotesize
\caption{Candidate attitudes towards policy promises \label{table_2}} \medskip
\estauto{"Tables/Table 2.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Mean of responses among incumbents, challengers, and other candidates to our candidate survey. \textit{Different promises}: believes that candidates make different promises on radio versus on-the-ground campaigning; \textit{Rally credibility}: believes that promises made by candidates at rallies are very likely to be fulfilled; \textit{Radio credibility}: believes that promises made by candidates on radio are very likely to be fulfilled. }
\end{threeparttable}
\end{table}

\begin{table}[!!htpb]\centering
\begin{threeparttable}\footnotesize
\caption{Reasons cited for debate participation \label{table_3}} \medskip
\estauto{"Tables/Table 3.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Mean of responses among incumbents, challengers and other candidates to our candidate survey. Candidates were allowed to cite more than one reason for participation. \textit{Duty}: cited democratic duty; \textit{Policies}: cited opportunity to present policy platform; \textit{Competence}: cited opportunity to demonstrate competence; \textit{Publicity}: cited opportunity for free campaign publicity; \textit{Radio}: cited the benefits of radio broadcasting reaching a large audience; \textit{Attack}: cited opportunity to attack other candidates. }
\end{threeparttable}
\end{table}

\begin{table}[!!htpb]\centering
\begin{threeparttable}\footnotesize
\caption{Transcript descriptive statistics \label{table_4}} \medskip
\estauto{"Tables/Table 4.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Mean values of variables among incumbents, challengers and other candidates based on our debate transcriptions.  \textit{Intro words}: number of words spoken in debate introduction; \textit{Education emphasis}: candidate highlighted their education in introduction; \textit{Experience emphasis}: candidate highlighted their experience in introduction; \textit{CSDF words}: number of words spoken about ways to improve management of County Social Development Funds; \textit{LSP words}: number of words spoken about priorities for spending Legislative Support Projects funds; \textit{Attacked}: candidate was verbally attacked by another candidate; \textit{Attacker}: candidate verbally attacked another candidate.}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Effects on candidate debate participation \label{table_5}}
\estauto{"Tables/Table 5.tex"}{8}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes: share of candidate types (all, incumbent, challenger, other candidate) who attended a debate out of all candidates in that district. Panel D has fewer observations due to only three candidates running in two districts (and hence no `other candidates' defined). See Table A26 for all predetermined covariate coefficients. \Spec}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable} \scriptsize
\caption{Effects on voting outcomes \label{table_6}}
\estauto{"Tables/Table 6.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes: Panel 1: respondent reported voting for incumbent (Panel 1.A.) or a challenger (Panel 1.B.) at endline; Panel 2: votes divided by registered voters for incumbent (Panel 2.A.) or challengers (Panel 2.B.) using polling station-level data. Interactions: Panel 1: Columns 4-6: standardized candidate-level measures of predicted debate performance; 7-9: standardized respondent-candidate-level measures of policy alignment (measured at baseline); Panel 2 uses district-level analogs of interaction terms (see Data section). See Tables A27-A30 for all predetermined covariate coefficients. \SpecMixed}
\end{threeparttable}
\end{table}

\input{"Tables/Table 7.tex"}

\clearpage

\setcounter{figure}{1}

\begin{figure}[!!htpb]
\begin{center}
\caption{Distribution of compliance strata by baseline candidate characteristics \label{fig_2}}
\includegraphics[scale=0.8]{"Figures/Figure 2.pdf"}
\end{center}
\footnotesize \textit{Notes}: Nonparametric estimation of compliance status across standardized values of baseline characteristics for incumbents (top) and challengers (bottom). At a given value of each baseline characteristic, the fitted value indicates the probability of a candidate type being an always-taker or complier. Optimal bandwidths computed following \citet{calonico2018effect}.
\end{figure}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Effects on debate exposure and information acquisition \label{table_8}}
\estauto{"Tables/Table 8.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes are standardized. Panel A: index of (1) indicator for respondent heard debate between baselind and endline; (2) how often respondent heard debate by endline; Panel B: index of (1) indicator for respondent's stated debate winner attended debate; (2) share of candidates respondent claims participated; (3) share of predicted leading candidates respondent claims participated; Panel C: change in how many questions about CSDF management respondents answered correctly. Panel D: index of (1) change in how much respondents listened to radio; (2) change in how much they discussed politics with friends; (3) how much they accessed other sources of political information. See Table A18 for disaggregated indicator-level results and Tables A31-A32 for predetermined covariate coefficients. \Spec}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Effects on updating about candidates \label{table_9}}
\estauto{"Tables/Table 9.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes are standardized. Panel 1: Columns 1-3: change in certainty about candidates' competence; 4-6: change in certainty about candidates' policy priorities; Panel 2: Columns 1-3: change in perceptions of candidates' competence; 4-6: change in correctly learning candidates' policy priorities. See Tables A33-A34 for predetermined covariate coefficients. \SpecCand}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Effects on campaigning \label{table_10}}
\estauto{"Tables/Table 10.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes are standardized. Columns 1-3: index of how often candidates (1) visited; (2) distributed leaflets; (3) bought votes in respondents' communities; 4-6: index of how often respondents heard candidates on radio in two weeks before election. See Table A19 for disaggregated indicator-level results, and Table A35 for predetermined covariate coefficients. \SpecCand}
\end{threeparttable}
\end{table}

\clearpage

\section{Supplementary materials}

\setcounter{table}{0}
\setcounter{figure}{0}
\renewcommand{\thetable}{A\arabic{table}}
\renewcommand{\thefigure}{A\arabic{figure}}

\subsection{Figures}

\begin{figure}[!!htpb]
\begin{center}
\caption{Scope conditions \label{fig_A1}}
\includegraphics[width=0.85\textwidth]{"Figures/Figure A1.pdf"}
\end{center}
\footnotesize \textit{Notes}: Figure presents region-level (top panel) and country-level within sub-Saharan Africa (bottom panel) indexes drawn from V-Dem data (v14) averaged over the period 2010-2020. \textit{Clientelism} index comes from \texttt{v2xnp\_client}; \textit{Party institutionalization} index comes from \texttt{v2xps\_party}; \textit{Free and fair elections} index comes from \texttt{v2xel\_frefair}.
\end{figure} 

\begin{figure}[!!htpb]
\begin{center}
\caption{Treatment effects of rebroadcasting intervention \label{fig_A2}}
\includegraphics[width=0.85\textwidth]{"Figures/Figure A2.pdf"} \\
\end{center}
\footnotesize \textit{Notes}: Treatment effect estimates of cross-randomized rebroadcasting intervention (see Appendix A.5) for main results in Tables 6-10. Estimated using specifications analogous to Equation (1) and all analysis of the intensive invitation intervention featured in the manuscript. 90\% and 95\% confidence intervals plotted, with colors indicating weights as defined in the manuscript.
\end{figure} 

\begin{figure}[!!htpb]
\begin{center}
\caption{Predicted versus observed performance among participating candidates \label{fig_A3}}
\includegraphics[width=0.75\textwidth]{"Figures/Figure A3.pdf"} \\
\end{center}
\footnotesize \textit{Notes}: Figure correlates observed versus predicted measures of debate performance across electoral districts assigned to control and treatment. See Appendix A.6.3 for additional details.
\end{figure}

\begin{figure}[!!htpb]
\begin{center}
\caption{Correlates of predicted performance measure\label{fig_A4}}
\includegraphics[width=0.8\textwidth]{"Figures/Figure A4.pdf"} \\
\end{center}
\footnotesize \textit{Notes:} Figure presents coefficients from a regression of our predicted debate performance measure onto a set of standardized candidate-level characteristics. 90\% and 95\% confidence intervals plotted. See Appendix A.6.3 for additional details. \normalsize
\end{figure}

\begin{figure}[!!htpb]
\begin{center}
\caption{Correlation between debate performance and campaigning in control districts \label{fig_A5}}
\includegraphics[width=0.75\textwidth]{"Figures/Figure A5.pdf"} \\
\end{center}
\footnotesize \textit{Notes}: Figure correlates standardized debate performance among candidates in districts assigned to control with the standardized campaigning outcomes (whether on-the-ground or radio) used in Table 10.
\end{figure}

\clearpage

\subsection{Tables}

\subsubsection{Descriptive statistics}

\begin{table}[!htbp]\centering
\begin{threeparttable}  \scriptsize
\caption{Descriptive statistics \label{table_A1}}
\estauto{"Tables/Table A1.tex"}{4}{crrrr}
\Figtext{\scriptsize \textit{Notes}: Descriptive statistics relating to all predetermined covariates used in the regression analyses. \textit{Sources:} Panel A: Debate variables from Internews. All 2017 and 2011 variables come from National Elections Commission (NEC). All 2008 variables come from 2008 Population and Housing Census. \textit{Share with GSM coverage} comes from Collins Mobile Coverage Explorer; \textit{Share owns a radio} and \textit{Share gets radio news often} come from Afrobarometer. \textit{Avg. N radio stations covering each town} comes from Internews. Pabel B: All come from researchers' panel survey. Panel C:  Radio station variables come from Internews. All other variables come from NEC. 90\% of polling stations were in locations where a polling place (Voter Registration Center, VRC) existed in 2011. For new polling stations we assign district-level averages.}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}  \footnotesize
\caption{Descriptive statistics (respondent-level outcomes) \label{table_A2}}
\estauto{"Tables/Table A2.tex"}{4}{crrrr}
\Figtext{\footnotesize \textit{Notes}: Descriptive statistics relating to all unstandardized, disaggregated outcomes variables used in the respondent-level regressions as outcome variables.}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable} \footnotesize
\caption{Descriptive statistics (respondent-candidate level outcomes) \label{table_A3}}
\estauto{"Tables/Table A3.tex"}{10}{crrrrrrrrrrrr}
\Figtext{\footnotesize \textit{Notes}: Descriptive statistics relating to all unstandardized, disaggregated outcomes variables used in the respondent-candidate level regressions as outcome variables.}
\end{threeparttable}
\end{table}

\clearpage

\subsubsection{Balance on predetermined covariates}

\setcounter{table}{4}

\begin{table}[!htbp]\centering\footnotesize
\begin{threeparttable}
\caption{District-level balance \label{table_A5}}
\estauto{"Tables/Table A5.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Descriptions of all variables can be found in Table \ref{table_A1}. \SpecBalance}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering\footnotesize
\begin{threeparttable}
\caption{Individual-level balance (district covariates) \label{table_A6}}
\estauto{"Tables/Table A6.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Descriptions of all variables can be found in Table \ref{table_A1}. \SpecBalance}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering\footnotesize
\begin{threeparttable}
\caption{Individual-level balance (individual covariates) \label{table_A7}}
\estauto{"Tables/Table A7.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: \SpecBalance}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering\footnotesize
\begin{threeparttable}
\caption{Polling station-level balance (district covariates) \label{table_A8}}
\estauto{"Tables/Table A8.tex"}{12}{c}
\Figtext{\footnotesize \textit{Notes}: Descriptions of all variables can be found in Table \ref{table_A1}. \SpecBalance}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering\scriptsize
\begin{threeparttable}
\caption{Polling station-level balance (PS covariates) \label{table_A9}}
\estauto{"Tables/Table A9.tex"}{12}{c}
\Figtext{\footnotesize \textit{Notes}: Descriptions of all variables can be found in Table \ref{table_A1}. Analysis throughout is at the polling station-level, where multiple polling stations exist within a single location called a VRC. 1780/2080 VRCs existed in the 2011 election; for these 300 new VRCs we assign district-level averages. \SpecBalance}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering\footnotesize
\begin{threeparttable}
\caption{Incumbent balance \label{table_A10}}
\estauto{"Tables/Table A10.tex"}{12}{c}
\Figtext{\footnotesize \textit{Notes}: Outcome variables are plenary session attendance measures taken from legislator scorecards for 2016. Legislators either attend, are absent, or are away from Monrovia for each plenary session. \SpecBalance}
\end{threeparttable}
\end{table}

\begin{landscape}
\begin{table}[!htbp]\centering\footnotesize
\begin{threeparttable}
\caption{Candidate-level balance \label{table_A11}}
\estauto{"Tables/Table A11.tex"}{12}{c}
\Figtext{\footnotesize \textit{Notes}: Panel A: unweighted balance tests for all candidate respondents, Panel B: restricted to incumbents; Panel C: restricted to predicted challengers. Outcomes: response to candidate survey; age in years; indicator for whether candidate completed university; indicator for whether they own or manage a radio station; indicator for candidate being male; indicator for candidate having run for office before; indicator for candidate having a government job before; indicator for candidate having advocacy experience; indicator for candidate working for an NGO before; indicator for belonging to ruling Unity Party; measure of policy alignment; measure of predicted debate performance. \SpecBalance}
\end{threeparttable}
\end{table}
\end{landscape}

\begin{landscape}
\begin{table}[!htbp]\centering\footnotesize
\begin{threeparttable}
\caption{Candidate-level balance (weighted) \label{table_A12}}
\estauto{"Tables/Table A12.tex"}{12}{c}
\Figtext{\footnotesize \textit{Notes}: Observations weighted by the inverse of the number of responding candidate types in a given district. Since there is one incumbent per district, for incumbents this would be identical to Panel B of Table A11 and thus we omit it. Panel A: all candidate respondents, Panel B: restricted to predicted challengers. Outcomes: response to candidate survey; age in years; indicator for whether candidate completed university; indicator for whether they own or manage a radio station; indicator for candidate being male; indicator for candidate having run for office before; indicator for candidate having a government job before; indicator for candidate having advocacy experience; indicator for candidate working for an NGO before; indicator for belonging to ruling Unity Party; measure of policy alignment; measure of predicted debate performance. \SpecBalance}
\end{threeparttable}
\end{table}
\end{landscape}

\clearpage

\subsubsection{Demography-weighted specifications}

\begin{table}[!htbp]\centering
\begin{threeparttable}\scriptsize
\caption{Candidate debate participation (demography-weighted) \label{table_A13}}
\estauto{"Tables/Table A13.tex"}{5}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes: share of candidate types (all, incumbent, challenger, other candidate) who attended a debate out of all candidates in that district. Panel D has fewer observations due to only three candidates running in two districts (and hence no `other candidates' defined). \SpecWgt}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable} \footnotesize
\caption{Voting outcomes (demography-weighted) \label{table_A14}}
\estauto{"Tables/Table A14.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes: respondent reported voting for incumbent (Panel 1.A.) or a challenger (Panel 1.B.) at endline. Interactions: Columns 4-6: standardized candidate-level measures of predicted debate performance; 7-9: standardized respondent-candidate-level measures of policy alignment (measured at baseline). \SpecWgt}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Debate exposure and information acquisition (demography-weighted) \label{table_A15}}
\estauto{"Tables/Table A15.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes are standardized. Panel A: index of (1) indicator for respondent heard debate between baselind and endline; (2) how often respondent heard debate by endline; Panel B: index of (1) indicator for respondent's stated debate winner attended debate; (2) share of candidates respondent claims participated; (3) share of predicted leading candidates respondent claims participated; Panel C: change in how many questions about CSDF management respondents answered correctly. Panel D: index of (1) change in how much respondents listened to radio; (2) change in how much they discussed politics with friends; (3) how much they accessed other sources of political information. See Table A18 for disaggregated indicator-level results and Tables A31-A32 for predetermined covariate coefficients. \SpecWgt} 
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Updating about candidates (demography-weighted) \label{table_A16}}
\estauto{"Tables/Table A16.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes are standardized. Panel 1: Columns 1-3: change in certainty about candidates' competence; 4-6: change in certainty about candidates' policy priorities; Panel 2: Columns 1-3: change in perceptions of candidates' competence; 4-6: change in correctly learning candidates' policy priorities. \SpecWgt}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Candidate campaigning (demography-weighted) \label{table_A17}}
\estauto{"Tables/Table A17.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes are standardized. Columns 1-3: index of how often candidates (1) visited; (2) distributed leaflets; (3) bought votes in respondents' communities; 4-6: index of how often respondents heard candidates on radio in two weeks before election. See Table A19 for disaggregated indicator-level results, and Table A35 for predetermined covariate coefficients. \SpecWgt}
\end{threeparttable}
\end{table}

\clearpage

\subsubsection{Effects on disaggregated outcomes}

\begin{table}[!htbp]\centering\tiny
\scalebox{0.9}{
\begin{threeparttable}
\caption{Debate exposure (supplementary) \label{table_A18}}
\estauto{"Tables/Table A18.tex"}{3}{c}
\Figtext{\scriptsize \textit{Notes}: All outcome variables are described in Appendix A.6. \Spec}
\end{threeparttable}
}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Candidate campaigning (supplementary) \label{table_A19}}
\estauto{"Tables/Table A19.tex"}{3}{c}
\Figtext{\footnotesize \textit{Notes}: All outcome variables are described in Appendix A.6. \SpecCand}
\end{threeparttable}
\end{table}

\clearpage

\subsubsection{Other tables}

\begin{table}[!htbp]\centering
\begin{threeparttable}  \scriptsize
\caption{Candidate debate participation (supplementary) \label{table_A20}}
\estauto{"Tables/Table A20.tex"}{3}{c}
\Figtext{\footnotesize \textit{Notes}: Outcome: share of the respective set of candidates (winner, actual challenger) who attended a debate out of all candidates in that district. Actual challengers are candidates who ranked in the top three in their race in the election but were not the incumbent. Actual other candidates are those who did not rank in the top three. Panel C has fewer observations due to only three candidates running in two districts (and hence no `actual other candidates' defined). \Spec}
\end{threeparttable}
\end{table}

\begin{table}[htbp]\centering
\begin{threeparttable}  \scriptsize
\caption{Candidate debate participation (district-level) \label{table_A21}}
\estauto{"Tables/Table A21.tex"}{3}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes: share of candidate types (all, incumbent, challenger, other candidate) who attended a debate out of all candidates in that district. Panel D has fewer observations due to only three candidates running in two districts (and hence no `other candidates' defined). For weighted specifications, \textit{Reg} is the number of registered voters in that district. \SpecBalance} 
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Debate rebroadcasting \label{table_A22}}
\estauto{"Tables/Table A22.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes: Panel A: number of rebroadcasts confirmed by radio monitors; Panel B: number of rebroadcasts based on survey of radio stations, including those not contracted to rebroadcast but recorded as being present in the debate venue. \Spec}
\end{threeparttable}
\end{table}

\begin{table}[htbp]\centering
\begin{threeparttable} \scriptsize
\caption{Effects on voting outcomes (Panel specification) \label{table_A23}}
\estauto{"Tables/Table A23.tex"}{10}{c}
\Figtext{\footnotesize \textit{Notes}: Panels 1.A. and 1.B: Outcome: Change in indicator that respondent reported voting for incumbent (Panel A) or a predicted challenger (Panel B) between baseline and endline. Columns 4-6 interact treatment assignment with standardized candidate-level measures of predicted debate performance; 7-9 interact treatment with standardized respondent-candidate-level measures of policy alignment (measured at baseline). \SpecMixedPanel}
\end{threeparttable}
\end{table}

\begin{table}[!htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Rebroadcasting intervention \label{table_A24}}
\estauto{"Tables/Table A24.tex"}{3}{c}
\Figtext{\footnotesize \textit{Notes}: Outcome is standardized index of debate exposure. Panel A uses respondents in full sample of electoral districts; Panel B restricts to respondents in electoral districts assigned to \textit{Intensive Invite}.  \Spec}
\end{threeparttable}
\end{table}

\begin{table}[htbp]\centering
\begin{threeparttable}\footnotesize
\caption{Effects on coordination \label{table_A25}}
\estauto{"Tables/Table A25.tex"}{3}{c}
\Figtext{\footnotesize \textit{Notes}: Outcomes: How much respondents discussed the debate with friends; How much this discussion led to coordinating their vote choices. \Spec} 
\end{threeparttable}
\end{table}

\clearpage

\subsubsection{Tables with predetermined covariate coefficients}


\begin{table}[!htbp]\centering\tiny
\scalebox{0.75}{
\begin{threeparttable}
\caption{Candidate debate participation (Table 5) \label{table_A26}}
\estauto{"Tables/Table A26.tex"}{15}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes: share of candidate types (all, incumbent, challenger, other candidate) who attended a debate out of all candidates in that district. Panel D has fewer observations due to only three candidates running in two districts (and hence no `other candidates' defined).  \Spec}
\end{threeparttable}
}
\end{table}

\begin{landscape}
\begin{table}[htbp]\centering
\begin{threeparttable} \tiny
\caption{Voting outcomes (Table 6, Panel 1.A.) \label{table_A27}}
\estauto{"Tables/Table A27.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes: respondent reported voting for incumbent at endline. Interactions: Columns 4-6: standardized candidate-level measures of predicted debate performance; 7-9: standardized respondent-candidate-level measures of policy alignment (measured at baseline). \SpecMixed}
\end{threeparttable}
\end{table}
\end{landscape}

\begin{landscape}
\begin{table}[htbp]\centering
\begin{threeparttable} \tiny
\caption{Voting outcomes (Table 6, Panel 1.B.) \label{table_A28}}
\estauto{"Tables/Table A28.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes: respondent reported voting for a challenger at endline. Interactions: Columns 4-6: standardized candidate-level measures of predicted debate performance; 7-9: standardized respondent-candidate-level measures of policy alignment (measured at baseline). \SpecMixed}
\end{threeparttable}
\end{table}
\end{landscape}

\begin{landscape}
\begin{table}[htbp]\centering
\begin{threeparttable} \tiny
\caption{Voting outcomes (Table 6, Panel 2.A.) \label{table_A29}}
\estauto{"Tables/Table A29.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes: Panel 2: votes divided by registered voters for incumbent using polling station-level data. Interactions: district-level analogs of interaction terms (see Data section and above tables). \SpecMixed}
\end{threeparttable}
\end{table}
\end{landscape}

\begin{landscape}
\begin{table}[htbp]\centering
\begin{threeparttable} \tiny
\caption{Voting outcomes (Table 6, Panel 2.B.) \label{table_A30}}
\estauto{"Tables/Table A30.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes: Panel 2: votes divided by registered voters for challengers using polling station-level data. Interactions: district-level analogs of interaction terms (see Data section and above tables). \SpecMixed}
\end{threeparttable}
\end{table}
\end{landscape}


\begin{table}[htbp]\centering\tiny 
\begin{threeparttable}
\caption{Debate exposure and information acquisition (Table 8, Panels A-B) \label{table_A31}}
\estauto{"Tables/Table A31.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes are standardized. Panel A: index of (1) indicator for respondent heard debate between baselind and endline; (2) how often respondent heard debate by endline; Panel B: index of (1) indicator for respondent's stated debate winner attended debate; (2) share of candidates respondent claims participated; (3) share of predicted leading candidates respondent claims participated. \Spec}
\end{threeparttable}
\end{table}

\begin{table}[htbp]\centering\tiny 
\begin{threeparttable}
\caption{Debate exposure and information acquisition (Table 8, Panels C-D) \label{table_A32}}
\estauto{"Tables/Table A32.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes are standardized. Outcomes: Panel C: change in how many questions about CSDF management respondents answered correctly. Panel D: index of (1) change in how much respondents listened to radio; (2) change in how much they discussed politics with friends; (3) how much they accessed other sources of political information. See Table A18 for disaggregated indicator-level results and Tables A31-A32 for predetermined covariate coefficients. \Spec}
\end{threeparttable}
\end{table}


\begin{table}[htbp]\centering\tiny
\begin{threeparttable}
\caption{Updating about candidates (Table 9, Panel 1) \label{table_A33}}
\estauto{"Tables/Table A33.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes are standardized. Outcomes: Columns 1-3: change in certainty about candidates' competence; 4-6: change in certainty about candidates' policy priorities. \SpecCand}  
\end{threeparttable}
\end{table}

\begin{table}[htbp]\centering\tiny
\begin{threeparttable}
\caption{Updating about candidates (Table 9, Panel 2) \label{table_A34}}
\estauto{"Tables/Table A34.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes are standardized. Outcomes: Columns 1-3: change in perceptions of candidates' competence; 4-6: change in correctly learning candidates' policy priorities. \SpecCand}  
\end{threeparttable}
\end{table}

\begin{table}[htbp]\centering\tiny
\begin{threeparttable}
\caption{Campaigning responses (Table 10) \label{table_A35}}
\estauto{"Tables/Table A35.tex"}{20}{c}
\Figtext{\scriptsize \textit{Notes}: Outcomes are standardized. Columns 1-3: index of how often candidates (1) visited (2) distributed leaflets (3) bought votes in respondents' communities; 4-6: index of how often respondents heard candidates on radio in two weeks before election. \SpecCand}
\end{threeparttable}
\end{table}

\clearpage

\bibliographystyle{apsr}
\bibliography{refs.bib}


\end{document}